INDEX

Explanations

elements related to code structure and object manipulation within programming

oai_token-act-pair · gpt-4o-mini Triggered by @bot

that European nations

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 didalam

-0.58

typeorm

-0.52

 nôtre

-0.49

mektedir

-0.48

spaceBetween

-0.46

 vôtre

-0.45

 defaultstate

-0.45

 étudi

-0.45

maktadır

-0.43

 wundersch

-0.43

POSITIVE LOGITS

:✨

0.52

 snippetHide

0.50

farin

0.46

newswire

0.45

聙

0.45

 ausp

0.44

[...,

0.42

aclk

0.41

出版年

0.41

qed

0.40

Activations Density 0.109%

elements related to code structure and object manipulation within programming

that European nations

No Comments

No Known Activations